COVID-19 and seasonality: A thread.

So, we're entering the typical season for respiratory viruses in the middle of a respiratory virus pandemic, which means we're going to get lots of explanations that feel intuitive.

Be careful with those. Lets talk flu... 1/8

#epitwitter

The first epidemiology projects I worked on were focused on influenza seasonality - something that was already well studied at the time. My very first paper? A review of the proposed casual mechanisms for seasonality. If you really want to read it: https://jvi.asm.org/content/81/11/5429

2/8

Epidemiology is *very* good at mathematically representing seasonality - we can make viral transmission parameters go up and down in nice, seasonal patterns, and get these patterns to fit the data.

This is called "seasonal forcing".

3/8

The tricky bit is, unlike a lot of modeling parameters, this one isn't liked to a causal mechanism. "Beta goes up sometimes, and down other times". We have *lots* of "Just So" stories as to why.

4/8

It's indoor crowding and heating. It's changes in relative humidity and ambient temperature. It's immune suppression due to reduced Vitamin D. It's the seasonal dynamics of schools.

A lot of these feel right. None of them are a comprehensive explanation.

5/8

Now we're not sure if COVID-19 will be seasonal. There's evidence ( @maiamajumder ) that it's not temperature dependent. That doesn't mean it isn't seasonal.

The pessimist in me says it probably is.

6/8

What this means is there will be a lot of intuitive feeling articles about how the seasonality of COVID-19 is driven by X, so we clearly need to do Y.

These will make sense. They will feel right

7/8

What they actually are is very strong causal arguments that we're not even all that secure in making for much, much better understood diseases.

Apply the appropriate amount of skepticism. Especially when they come paired with easy policy recommendations.

~fin~